[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [JSR 166 (Concurrency utilities)]

> The JSR says (section 2.5)
> > However, there are a few critical JVM/language enhancements 
> > surrounding atomicity and monitors necessary to obtain efficient 
> > and correct semantics. 
> I am surprised that this didn't kill the proposal, especially the
> JVM bit... Can you comment on the likelihood that these enhancements
> will be made  ...

There was not a single dissenting vote. We have the active cooperation
of JVM folks from Sun, and hopefully from IBM.  If you read through the
aims & scope page I listed, you'll see that we are only proposing JVM
changes that just about anyone would agree are easy, effective, and
long overdue: support for atomic instructions, trylocks, and multiple
condition variables, all of which are readily implementable on the
vast majority of platforms, and not too hard to fake on others. All
the rest will be in Java.

So I expect the native support to come through without problem for
JDK1.5. (Note however, that 1.5 could be a long time coming.)